According to Bloomberg, over the past several months, many leading artificial intelligence (AI) developers in the US have been racing to develop sophisticated AI agents that can perform complex tasks on behalf of users.

Now, a little-known Chinese startup claims to have jumped to the top.

Last week, Manus AI launched a beta version of what it calls a general AI agent, capable of screening profiles, planning travel itineraries, and analyzing stocks based on basic instructions from users.

The company says its service performs better in some ways than OpenAI's Deep Research, another recently released AI agent.

Manus co-founder and chief scientist Quy Dat Sieu said that while some AI agents require some level of guidance and supervision, their product is "truly autonomous."

An eye-catching introductory video from the company quickly went viral, sending people scrambling to secure a limited number of invitations to test the tool.

The initial buzz has drawn comparisons to DeepSeek, the Chinese startup that shocked Silicon Valley in January when it released a competing AI model it claimed to have developed at a fraction of what its American rivals spent on its technology.

Like DeepSeek, Manus is once again raising questions about America's leadership in artificial intelligence — this time in a product category that American tech companies see as a key investment area.

However, so far, the initial response from Manus users has been mixed. Professor Derya Unutmaz at the Jackson Laboratory, which studies cancer immunotherapy, praised the tool for its “good quality results” despite “taking longer to process tasks than OpenAI’s Deep Research.”

Others complained that the service was too slow and sometimes crashed before completing a task, possibly due to the company's limited computing resources. Some users also found it to be riddled with bugs.

“Manus is really an unfinished product,” said Yiran Chen, a professor of electrical and computer engineering at Duke University who has tested the service. The startup may hope that “by being the first mover, they can attract investors, despite the fact that the product is not fully developed.”

There are many unknowns about the company and its product, including the extent to which it is simply refining and building on existing models from other AI developers rather than building its own advanced systems—the latter approach could cost tens of millions of dollars or more.

According to reports from several Chinese news agencies, the company behind Manus, Butterfly Effect, has raised more than $10 million in investment capital.

Unlike DeepSeek and some US companies, Manus does not publish detailed papers alongside its products to provide information about how the technology was developed.

They also do not publicly release any source code or weights related to their AI for people to use independently.

In recent months, OpenAI, Anthropic, and other AI companies have released agents that can use a user’s computer to browse the internet, conduct online research, and complete various multi-step tasks. But some early users of Manus say the tool is no slouch compared to what’s already on the market.

“I have tried hundreds of AI tools,” said Ashutosh Shrivastava, a software developer in Bangalore who has used the service to build a website and a game. “I keep trying new ones every day. I have never seen anything like Manus.”

Source: https://www.vietnamplus.vn/manus-cua-trung-quoc-thach-thuc-my-trong-cuoc-dua-tac-nhan-ai-post1019822.vnp